http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
백동민,조현민 제어·로봇·시스템학회 2024 제어·로봇·시스템학회 논문지 Vol.30 No.1
. This paper presents a method for microcontrollers control using a recurrent neural network-based inverse model. Limited computational power in microcontrollers makes applying complex neural network structures challenging. We use the Elman network with a simple structure as an inverse model to address this issue. Elman network was used to model nonlinear control systems. The proposed method constructs a recurrent neural network-based inverse model in parallel to enhance the performance of the PID controller. The recurrent neural network uses the output generated by the PID controller as the past control input and compensates for the control inputs generated by the PID controller. We applied the proposed controller to a DC motor current control system and compared its performance with the PID controller that uses a deep neural network as an inverse model. We evaluated the control performance by applying a sine wave. The results show that the proposed controller has better tracking performance at 1, 3, and 5 Hz than the other controllers.